►
From YouTube: Kubernetes SIG CLI 20230531
Description
Kubernetes SIG CLI bi-weekly meeting on May 31st, 2023.
Agenda and Notes: https://docs.google.com/document/d/1r0YElcXt6G5mOWxwZiXgGu_X6he3F--wKwg-9UBc29I/edit#bookmark=kix.di46kpegaz22
A
Hello,
my
name
is
Sean
Sullivan
I'm,
your
moderator
for
the
May
31st
version
of
the
six
CLI
bi-weekly
meeting,
I'm
gonna
jump
directly
into
the
announcements,
we're
right
in
the
heart
of
the
128
version
or
sorry,
the
128
release,
and
so
we
have
next
week
the
production,
Readiness
review
and
a
week
after
that.
A
So
within
two
weeks
we
have
the
enhancement
freeze,
and
so
we
have
a
couple
of
caps
that
are
that
are
that
relate
to
the
6cli
and
we're
going
to
actually
have
a
few
words
on
those
as
we
progress
and
it's
about
I.
Think
four
to
five
weeks
after
that
in
July
18th,
that
we
have
the
code
freeze
and
another
month
after
that,
as
is
typical
for
the
release
itself,
so
we're
right
in
the
thick
of
it
right
now
and
there's
a
couple
other
announcements
on
the
six
CLI
leadership.
A
So
congratulations
to
Eddie
and
Natasha,
who
have
been
elevated
to
positions
for
Eddie
as
the
tech
lead
and
Natasha
as
a
new
chair,
I'm
stepping
down
as
chair.
Congratulations.
Thank
you
very
much
to
Sean
for.
A
That's
it's
been
an
honor
working
with
you
all
and
I
hope
to
continue
working
with
you
all,
and
so
let's
move
on
to
the
the
new
members
or
attendees,
and
so,
if
you
are
new
to
this
meeting
or
you
haven't
been
here
for
a
long
time
and
you
would
like
to
be
introduced
to
your
six
CLI
colleagues,
this
is
your
opportunity
to
to
say
hi
and
to
introduce
yourself.
Is
there
anybody
who
would
like
to
introduce
themselves.
C
Just
about
everybody
I
see
here,
I've
I
recognize.
C
A
If
we
don't
have
any
introductions
of
new
members,
let's
move
on
to
the
sub
project
and
kept
updates.
The
first
is
the
the
transition
from
Speedy
to
websockets,
kep
I'll,
say
a
few
words
on
this
I've
gotten
a
lot
of
good
feedback
thanks
much
a
for
their
recent
feedback.
I'll
be
addressing
that
today.
A
A
There's
actually
been
I,
think
four
or
five
people
have
given
really
good
feedback
and
the
only
person
that
I
think
needs
to
give
feedback
who
hasn't
is
is
David
Eads
and
so
I'll
scroll
through
a
little
bit
of
this,
including
this
particular
diagram,
which
describes
fairly
well
at
a
high
level
what's
going
on
and
what
we
hope
to
do
is
put
in
a
another
proxy
on
the
API
server.
That's
going
to
translate
from
web
websockets
to
Speedy.
A
We
actually
have
it
implemented
already
in
a
proof
of
concept
PR
and
have
some
stress
tests,
which
is
beat
it
up.
Pretty
good,
so
it
looks
pretty
solid
right
now.
All
the
end-to-head
tests
were
working
on
that
proof
of
concept,
PR
I'll
move
on.
Does
anybody
have
any
questions
about
that
before
I?
Go
on
to
artists.
B
B
B
B
B
I
would
and
again
it
would
be
good
to
call
it
out
immediately
and
then
pointed
out
that
this
will
be
resolved
during
the
second
or
third
Alpha.
It
doesn't
matter
how
long
it'll
take
us
to
get
there,
but
at
least
before
we
do
a
final
GA
of
this
functionality.
B
We
need
to
make
sure
that
we
have
all
of
them
migrated,
but
other
than
that
I
think
there
are
a
couple
of
holes,
but
you're,
probably
aware
because
you
you
put
it
on
the
result
in
in
a
couple
of
them,
but
you
said
you're
gonna
be
addressing
all
of
those
later
today,
so
yeah
just
try
to
get
it
implement,
it
will
stay
in.
B
We
will
be
iterating
on
on
that
we've
seen
in
the
past,
a
couple
of
people
already
tried
it
and
did
not
succeeded,
which
I'm
expecting
it'll
just
take
a
sufficient
amount
of
time,
at
least
a
couple
Alpha.
Maybe
couple
betas
before
we
will
do
the
the
final
ga.
A
The
yeah
yeah
thanks
for
that,
thanks
for
that
feedback
and
yeah
just
to
to
continue
on
the
port
forward.
So
there's
the
kep
and
the
proof
of
concept.
A
Pr
are
for
the
remote
command
sub
protocol
in
and
not
the
port
forward
sub
protocol,
and
we
did
that
on
purpose
because
the
because
we
just
needed
to
get
the
scope
small
enough
that
we
could
make
some
progress,
because
this
has
been
going
on
for
eight
years
and
and
actually
Clayton
had
tried
getting
port
forward
onto
websockets
back
in
2017,
and
he
realized
that
it
was
much
bigger
than
what
he
had
initially
imagined.
A
And
so,
instead
of
trying
to
do
too
much
and
not
getting
anything
done,
we
figured
let's
try
to
just
get
three
of
these
four
commands
done,
where
they're
going
to
be
just
for
one
leg
from
the
client
from
Coupe
control
to
the
API
server
will
be
using
websockets,
and
that
gives
us
95
99
of
what
we
need,
because
what
we're
trying
to
do
is
solve
the
proxies
or
gateways
or
any
intermediary
that
doesn't
speak,
speedy
and
and
they
don't
none
of
them
do
anymore,
because
it's
been
deprecated
for
eight
years,
we're
going
to
move
it
to
websockets,
which
would
you
know
basically
allow
any
any
of
these
particular
intermediaries
to
use
a
modern
streaming
protocol,
and
we
we
get.
A
You
know
like
I,
said
95
of
what
we
need
and
then
it
will
continue
to
be
a
speedy
from
the
API
server
down
to
the
kublet
and
down
to
the
container
runtime
and,
and
what
that
gives
us
is
that
we
don't
have
to
actually
change
multiple
run
times,
because
the
servers
are
on
these
runtimes
and-
and
so
this
actually
has
massively
scoped
this
down
to
something
that
is
achievable
with
you
know.
For
for
us,
these
three
commands
from
Coupe
control
to
to
the
API
server.
So,
okay,
so
that's
yeah.
A
Thanks
for
that
feedback
and
as
I
said
I'll
be
getting
to
that
today
and
throughout
probably
the
next
couple
days
to
make
sure
that
you
know
it's
in
good
shape
for
this
upcoming
Alpha
marked
as
implementable
and
so
there's
also
Arda
is
doing
a
an
enhancement
and
I
have
it
here
and
Ardo.
Would
you
like
to
talk
about
this
particular
account?
Hi.
D
This
one
is
relatively
small
cap
comparing
to
the
websocket
migration,
but
I
put
it
in
the
agenda
just
for
Rising
the
awareness
we
are
doing,
adding
a
new
flag
or
interactivity
delete
and
I
suppose
it
is
requested
widely
and
yeah
feel
free
review
and
drop.
Your
comments.
A
That
includes
a
deal
I
hope
I
said
pronounce
your
name
correctly.
E
A
Okay
yeah,
so
there
have
been
a
lot
of
drain
issues
is,
is
Eddie
here
I
remember
he.
B
No,
it's
no
sorry
I
just
put
food
in
my
mouth.
It's.
A
Can
we
move
on
to
leonids
question.
F
Yeah
so
I
have
a
short
question.
I
was
looking
for
samples
and
didn't
find
them
I'm
working
on
some
implementation
changes
for
the
behavior
of
one
of
the
commons
and
they
pay
attention
that
a
certain
logic
about
parsing,
flag
values
is
not
tested
and
the
change
actually
broke
it
and
without
solo
code
review
this
would
be
missed.
F
So
my
question
is:
what
is
the
right
way
to
test
such
a
logic,
because
it
is
done
in
some
like
internal
method
for
the
code,
so
should
I
just
add
the
plane,
straightforward
method
for
testing
I
I've
commented
so
a
weight
test
go
it
tests
kind
of.
F
It
assumes
it
is
already
working,
so
he
he
decent
to
end
from
the
moment
that
the
correct
parameters
are
executed
and
the
command
runs
to
the
to
the
expected
results
and
I'm
looking
for
the
right
way
to
test
the
first
part
when
the
common
just
invoked
from
the
command
line
and
to
the
moment
the
comment
logic
is
executed,
I
hope
I
explained
it
in
this
explanation
makes
sense
to
you.
B
So,
basically,
you
have
two
options
for
how
to
verify
this,
that
I
answered
correctly,
but
the
unit
tests
that
are
living
in
way
test
ago.
They
will
be
verifying
that
the
value
that
is
already
parsed
initially
by
the
by
Kroger
Library,
that
the
value
is
reasonable
and
that
it
may
be
sense
and
you
can
either
run
validate
or
complete,
doesn't
matter.
It
will
depend
on
what
you're
expecting
to
happen
in
the
particular
slide
value
in
the
in
one
of
these
methods
and
then
eventually
run
to
get
the
expected
results.
B
But
that
assumes
that
your
the
value
that
you
have
and
in
the
weight
options
structure
is
the
one
that
you
expected
to
be
parsed
from
the
user
input.
If,
on
the
other
hand,
you
want
to,
if
you
want
to
verify
the
corporate
parsing,
because
for
whatever
reason,
there's
some
additional
parsing
requirements,
whether
that's
batch,
related
or
program,
related
things.
B
For
those
cases,
there
are
it's
hard
to
call
integration
tests,
but
they
are
and
and
to
a
certain
degree,
there's
a
bunch
of
bash
like
Chris
that
exists
under
test.
Commerce
directories.
B
There
you'll
find
all
those
various
commands
which
are
being
executed
and
then
there's
there's
some
expectation
around
the
output
of
any
particular
command
which
are
being
tested
against
a
local
Cube
API
server
I
pasted
a
link
to
specifically
weight
sh
test,
which
has
some
preliminary
testing
for
the
weight
command
so
now,
depending
on
which
part
you're
targeting
you
will
want
to
write
either
a
test
integration,
command
tests
or
or
the
unit
test,
but
like
it
explains
the
unit
doesn't
verify
the
the
Cobra
parsing.
B
So
whatever
you
have
add
a
dash
dash
mic
flax
and
there
is
a
chance
that
there
will
be
some
kind
of
I.
Don't
know
bash
interactions
or
something
like
that.
That's
where
you
would
want
to
print
them
over
to
the
test
command.
F
So
what
I'm
looking
for
is
kind
of
me
there
of
these
two
cases,
I'm
the
command
itself.
It
receives
the
parsing
result
and
then
it
fills
the
process,
the
the
received
values
for
additional
parsing
of
the
arguments,
value
itself,
and
it
validates
this
argument.
Value
specifically,
is
speaking
Json
condition,
Json
path,
condition
is
formed
correctly
and
while
some
of
them
I'm
not
interested
to
parse
because
like
for
example,
Json
pass
expression,
get
parsed
by
the
jsonpass
parser,
so
I'm
not
going
to
implement
another
test
for
the
same
method.
F
However,
there
are
some
additional
expectations
like
the
construct
of
expression
equal
to
some
value,
and
the
behavior
of
this
equal
differs
depending
on
how
the
the
this
string
is
provided,
and
this
part
is
kind
of
missing.
F
So
would
it
be
okay,
just
to
add
another
method,
intubate
underscore
test
and
to
call
the
parsing
private
parsing
function
inside
weight
directly
or
there
is
a
different
way
that
is
already
implemented
somewhere.
I,
don't
know.
B
I'm
just
looking
through
the
code
for
the
wave
command
and,
if
I
understand
correctly
you're.
Basically,
the
actual
parsing
of
the
Json
function
in
the
way
in
the
four
conditional
weight
is
happening
in
the
two
options
method.
So
in
the
in
your
weight
test
commands
you
would
want
to
get
from
or
manually
invoke
the
condition
function,
Bank,
4
or
directly
invoke
the
two
options.
B
Yeah,
so
that's
totally
fine
to
add
a
an
explicit
test
for
for
your
particular
use
cases.
It
can
be
either
invoking
two
options
or
you
can
directly
invoke
the
condition
function
for
which
will
and
then
invoke
the
the
actual
condition
function.
You're
going
to
get
that's
that.
That's
totally
reasonable
approach.
F
Okay,
thanks
and
then
I
can
convert
this
from
draft
to
the
active
PR
and
it
should
trigger
join
the
queue
of
the
for
the
reviewers
right
from
the
perspective
process
perspective
correct.
A
Thank
you
whoops,
so
that
that
was
the
extent
of
our
agenda
and
is
there
anything
else
that
we
would
like
to
talk
about
before
we
get
some
time
back
this
morning
or.
A
Okay,
well
thanks
for
joining
us
and
we'll
have
another
meeting
in
two
weeks.